Astron. Astrophys. 358, 521-534 (2000)

4. The galactic disk traced by embedded OB stars and by molecular clouds

4.1. Centroid and thickness of the disk

The formation of OB stars in the galactic disk is distributed in a layer coincident but thinner than that of molecular clouds as traced by CO. Fig. 9 shows the radial variations of the centroid and the thickness (HWHM) of the embedded massive star layer for the southern and northern Galaxy, compared with their H₂ counterparts. The molecular gas parameters were derived from the Columbia U. - U. Chile deep CO surveys of the I and IV galactic quadrants (Paper I; Cohen et al. 1986); from surveys of the Carina region by Grabelsky et al. (1987) and of the III galactic quadrant by May et al. (1988, 1993), analyzed by Casassus (1995); and from a composite CO dataset of the II quadrant analyzed by Digel (1991).

Fig. 9. The azimuthally averaged thickness (HWHM) (top) and the centroid (bottom) of the massive star formation layer (solid line), as a function of R, are compared to those of H2 (dashed line), as derived from the Columbia U. - U. Chile deep CO surveys of the Galaxy

The average thickness of the embedded massive star distribution, within the solar circle, is 73 pc (FWHM), compared with 118 pc (FWHM) for the layer of molecular gas traced by CO (Tables 6 and 7). It has been shown by Dame & Thaddeus (1994), however, that there are possibly two components for the molecular gas layer: a thin one, with a FWHM of 80 pc, similar to the thickness of the young massive stars layer derived here, and a thick and diffuse one, with a FWHM in excess of 500 pc.

The warping of the disk outside the solar circle, as traced by the centroid of both the massive star formation and the H₂ layer, appears to be more pronounced in the northern Galaxy than in the south. A tentative explanation is that the spiral arms outside the solar circle can be traced within a larger longitude range in the south than in the north; we may be therefore averaging out the warp in azimuth in the south because of the large longitude span (Casassus 1995; May et al. 1997). For example, it is apparent from Fig. 2 than there are many more sources detected outside the solar circle in the IV quadrant than in the I quadrant of the Galaxy.

The flaring of the molecular gas layer in the outer Galaxy is evidenced by both the molecular and by the embedded massive star distribution in the south. However, while the thickness of the massive star layer in the northern outer Galaxy grows like that in the south, the mean thickness of the molecular hydrogen layer in the northern outer Galaxy is only about half that of its southern counterpart. A possible explanation is that while the longitude coverage of our survey of embedded massive stars is equally complete north and south, the composite CO data set of the northern outer Galaxy analyzed by Digel (1991) only covers the region , compared to the complete longitude coverage of the III and IV quadrant CO surveys used by Casassus (1995) in his analysis. Therefore the azimuthal range for the average is much larger in the south than in the north; averaging the galactic warp over large scales would increase the mean thickness derived for the molecular gas layer.

4.2. Embedded OB stars, FIR surface luminosity, and H₂ surface density

The mean radial distribution for the 748 massive star forming regions analyzed here (Fig. 3) is maximum at , falls sharply toward the galactic center, and decays exponentially toward the outer Galaxy (Fig. 10). Of these regions, 492 () are found at , and 332 () at , the region of the molecular annulus (Paper I). The fraction of 44% grows to when the contamination of the sample by low luminosity regions in the solar neighborhood is considered (Sect. 3.5). A similar annular radial dependence for the number surface density of OB associations in the inner Galaxy has been found by Mc Kee & Williams (1997) using an older compilation of radio H II regions (Smith et al. 1978) binned in rings 2 kpc wide. Such annular shape was found also by Kent et al. (1994) for the NIR emission attributed to M giants, based on large scale satellite observations.

Fig. 10. From top to bottom: azimuthally averaged face-on FIR surface luminosity, as a function of R, originated by embedded massive stars (solid line) for the northern Galaxy (); the southern Galaxy (); and the whole Galaxy, compared with molecular hydrogen face-on surface density (short dashed line), as derived from the Columbia U. - U. Chile deep CO surveys of the Galaxy. Data are uncorrected for the expansion velocity field at the 3-kpc arm locus (Paper I)

The FIR surface luminosity produced by embedded massive stars for the Galaxy as a whole is even more restricted to the molecular annulus than their number density. Integration over the whole galactic disk yields a total FIR luminosity of . Within the solar circle the FIR luminosity is , 81% of the total, and for , the FIR luminosity is , representing of the total. The radial dependance of the FIR surface luminosity produced by embedded massive stars has also a narrower maximum than the H₂ surface density. A gaussian fit to the azimuthally averaged FIR surface luminosity, within the solar circle, yields a FWHM of (Table 4), % the FWHM obtained for the H₂ radial distribution, (Table 5). Down from the maximum, toward the outer Galaxy, the FIR surface luminosity decreases, as a function of R, much faster than the H₂ surface density, with an exponential scale length of , compared to for the H₂.

Table 4. Azimuthally averaged face-on FIR surface luminosity produced by regions of massive star formation

Table 5. Azimuthally averaged face-on surface density of molecular hydrogen

Table 6. Separate axisymmetric models for the FIR luminosity from UC H II regions: - 1.0

Table 7. Separate axisymmetric models for the H₂ distribution as traced by CO emission: - 1.0

The ratio of FIR luminosity produced by embedded OB stars to H₂ mass, which may be used as a large scale tracer of massive star formation efficiency, is not constant throughout the Galaxy and has also a peak at . Integrating the H₂ surface density, as traced by the Columbia U. - U. de Chile CO surveys of the Galaxy, we compute a total H₂ mass of for , and of for . For all the CO data we have used main beam antenna temperatures , with a main beam efficiency of 0.82 (Paper I); CO to H₂ conversion factor derived using the same CO data and EGRET gamma-ray observations by Hunter et al. (1997); a distance of the Sun to the galactic center of ; and no correction for helium abundance. The average FIR luminosity to H₂ mass ratio, within the solar circle, is ; for the ratio grows to ; and for there is a maximum of . Outside the solar circle, for , the derived mean ratio is ; this figure can be considered as an upper limit, since the conversion factor has been argued to be from 2 to 4 times larger in the outer Galaxy (May et al. 1997).

The radial distribution for the complete set of OB star formation candidate sites from WC89 has been derived by Comerón & Torra (1996), assuming axial symmetry but without the use of any kinematic information. Our present results do not confirm the peak in massive star formation that Comerón & Torra (1996) find at kpc, although it is fair to say that our analysis is oriented to the galactic disk, and that we have therefore cut out the region within of the galactic center. However, we detect only 1 source in the range , where we sampled 50% of the face-on area, while for their result to hold there should be more than 50 OB star forming regions in the area we sampled.

4.3. Large scale asymmetries within the solar circle

Among the most striking results of the early CO surveys of the Galaxy were the large scale north-south deviations from axial symmetry obtained from separate fits of axisymmetric models to the CO data (Robinson et al. 1983; Cohen et al. 1985). The maximum molecular gas face-on surface density in the molecular annulus is a factor of 1.47 higher in the I galactic quadrant than in the IV quadrant (Paper I), where the maximum is really a plateau extending roughly for . Because massive star formation occurs mostly in giant molecular clouds, it would be expected to find a similar trend in the FIR surface luminosity produced by embedded massive stars. What we have found, however, is an enhancement of massive star formation in the southern Galaxy within the solar circle.

A fraction of 58% of the FIR luminosity produced by embedded massive stars, within the solar circle, comes from the southern Galaxy - while the total H₂ mass, as traced by CO, is about the same north and south. A possible explanation of such result could be that massive stars in the I galactic quadrant are preferentially in the far side of the Galaxy, and/or that in the IV quadrant they are preferentially in the near side. Thus, by assuming axial symmetry in our models, we could be artificially inflating the luminosity in the south or lowering it in the north. However, because the FIR luminous dense molecular gas cores we have sampled in the CS(2-1) are well correlated in space and velocity with the population of giant molecular clouds in the Galaxy as traced by CO (Bronfman 1992), the same asymmetry should be found in the distribution of molecular gas. But what we find here for the FIR surface luminosity distribution is an asymmetry from north to south which is opposite in sign to that of the molecular gas surface density.

Massive star formation per unit H₂ mass, as a consequence, is higher in the IV than in the I galactic quadrant. The average FIR luminosity to H₂ mass ratio in the southern Galaxy, for , is ; for grows to 0.28; and for we find a galactic maximum of . In comparison, the corresponding numbers in the northern Galaxy are , , and . The enhancement of massive star formation at in the southern Galaxy appears to be correlated with the position of the Norma spiral arm tangent, where a large fraction of the most luminous sources in the sample are found. Massive star formation enhancements of this kind have been observed in the spiral arms of external galaxies, like for instance in M51 (Nakai et al. 1994), although there the starburst seems to be driven by a companion galaxy. In our own galaxy such burst could be driven by a bar - that would have to be large enough to produce an effect at almost 5 kpc from the galactic center.

4.4. The contribution of embedded OB stars to the FIR luminosity of GMCs in the Galaxy

Preliminary results of our ongoing analysis of GMCs harbouring embedded UC H II regions in the southern Galaxy show a fairly linear relation between the FIR luminosity from embedded OB stars and the virial mass of the clouds, with a mean FIR luminosity to mass ratio of (Bronfman et al. 2000). This ratio grows to , for GMCs associated with bright H II regions, when the FIR luminosity is computed from IRAS sky-flux images (Mooney & Solomon 1988). Therefore we suggest a lower limit of for the contribution of embedded OB stars to the total FIR emission from GMCs undergoing massive star formation. Our estimate can be considered as a lower limit because: (a) some embedded OB stars are likely to illuminate dust regions so large in angular size that they would not be identified as IRAS point sources and therefore not included in our sample and, (b) if identified, some of the source FIR flux may have been lost because of the background subtraction.

The ratio of FIR luminosity produced by embedded massive stars to virial mass for GMCs in the southern Galaxy, , is similar to the FIR surface luminosity to H₂ surface density ratio found at the peak of the southern distribution of embedded OB stars, . Therefore it is possible to suggest that most GMCs in that region of the Galaxy are presently forming massive stars, and it appears reasonable to use such value as a standard scale to evaluate the massive star formation efficiency per unit H₂ mass in the rest of the Galaxy relative to its maximum value. It is worth keeping in mind, though, that here we are averaging the H₂ surface density over large areas of the Galaxy, several kpc² in size. Therefore our maximum value for the FIR luminosity to H₂ mass ratio will be much smaller than that obtained when computed only for the neighborhood of a massive star forming region, with a typical FIR luminosity of and a virial mass of (Plume et al. 1997)

The total FIR luminosity of derived here for the embedded massive star layer, for , represents of the FIR luminosity assigned to the H₂ layer by Bloemen et al. (1990) from IRAS data, , and of the FIR luminosity derived from DIRBE data () by Sodrosky et al. (1997). Because in their respective analyses they consider all the FIR emission from the galactic plane, while here we compute only the FIR emission produced by dust fairly close to the heating stars, a strict lower limit of can be set for the contribution of embedded massive stars to the total FIR output from molecular gas within the solar circle.

© European Southern Observatory (ESO) 2000

Online publication: June 8, 2000
helpdesk.link@springer.de